News

Strange But True: Artificial Intelligence

Merriam-Webster defines artificial intelligence as “the capability of a machine to imitate intelligent human behavior.”  This term, which originated in 1956, represents a scientific field that is simultaneously fascinating and worrying. Researchers have made great strides in this field, enabling AI to have a role in music, entertainment, medicine, and several other areas of our society.  However, with these incredible advancements, people have posed questions as to the downsides of AI and how it will change human interaction in various industries. When examining this complicated and quickly evolving field of technology, then, there are many factors to analyze.

Artificial intelligence essentially encompasses different types of machine learning with the intent to make specific industries more efficient. Our ability as humans to quickly adapt to new information and adjust our actions accordingly is a valuable skill, and scientists want to find ways to mimic this process in machines. There are two branches of AI research: generalized AI and applied AI. The former creates machines that use intelligence to carry out a variety of tasks, while the latter focuses on developing artificial intelligence for a specific purpose for that field of research. Both of these fields have yielded incredible results over the past several years by creating computer programs that learn from sets of rules and commands and complete a given process within those parameters.

Many incredible technologies have been developed through the use of artificial intelligence. Recently, a research group created a trash can that can detect different types of food waste. Through the use of a camera and scales, the technology can read what has been thrown away to create metrics that can help restaurants recognize and reduce wastes of food.  At the Institute of Cancer Research in London, they have developed AI that can predict the spread of several types of cancer, allowing doctors to form possible treatment plans earlier than they could have using other technology. These are just a few of the countless programs that have been developed in an attempt to tackle issues plaguing the world today.

A scan and analysis of food waste completed by the AI trash can.

There are still several obstacles to overcome in artificial intelligence development. As artificial intelligence has to do with machines, there are often ways to override the system’s programming. In one study, a colorful, distorted pattern was created to fool image recognition software. Strangely, the pattern forces the AI to ignore all other items in its field of sight and only register this pattern, which it interprets as a toaster.  This bug does not present itself in every image recognition program, but it could have a significant impact on future technology. These patterns could cause issues with things like self-driving cars and security systems. For example, if the government adds AI-powered luggage screening software into airport security machinery, someone could put this type of pattern on a weapon to disguise it in their bag and smuggle that item onto their flight.

The imaging software’s interpretation of a banana and the warped pattern

While AI can take over some aspects of human processes and interactions, there are certain boundaries that are more difficult to push. Since artificial intelligence is based in machines and developed using logic, many of the more ambiguous forms of expression such as humor and emotions are unclear to computers. As an example, let’s consider last month’s humorous announcement of a new regime at clay, entitled “We ‘May’ Have Killed the S.E.s.”  We know this entire article was a joke–none of the authors would ever actually think of killing the Senior Editors–but for the sake of the narrative and the April Fools prank, that was the theme they presented.  However, how would we properly convey the reasoning behind this joke to a robot? If someone taught an artificially intelligent machine that the given article is funny, that could possibly turn into an understanding that murder in general is humorous.  Researchers are trying to incorporate an understanding of humor into AI, but since they have to teach via sets of rules and guidelines, it is quite difficult. As explained by Julia Rayz, an associate professor at Purdue University’s Department of Computer and Information Technology, “…When there are no clear rules – and there are no clear rules in human communication – what are we going to tell the computer to do, find rules that don’t exist?”

Surely, artificial intelligence will continue to improve in coming years.  There are many existing benefits to AI development, and it is interesting to ponder the new ideas researchers could execute through the use of this technology.  However, it is important to note the importance of human interaction within our society. No matter how smart computers become, we as people can experience and utilize methods of expressions that machines will never organically follow. God created us in His image, and we are the way we are because He is our Creator. While developers can strive to perfect AI machinery, no product of human innovation will ever come close to what God has made.

Sources:

https://interestingengineering.com/scientist-attempts-to-teach-ai-what-is-funny

https://www.forbes.com/sites/charlestowersclark/2019/01/28/making-ai-more-emotional-part-one/#6b0d52195fc1

https://www.forbes.com/sites/bernardmarr/2017/04/25/the-complete-beginners-guide-to-artificial-intelligence/#4c9865524a83

https://www.independent.co.uk/news/business/news/ai-bin-launched-food-waste-restaurants-winnow-vision-a8835736.html

https://www.independent.co.uk/news/health/cancer-artificial-intelligence-predict-progress-evolve-revolver-dna-mutation-genetic-changes-a8517631.html

http://www.copyright.com/blog/does-artificial-intelligence-signal-the-end-for-corporate-librarians-qa/

https://www.theverge.com/2018/1/3/16844842/ai-computer-vision-trick-adversarial-patches-google

https://www.bbc.com/news/technology-42554735

7 Comments

  1. A trash can that can think? ( almost) WOW!!

  2. Fascinating! Great job!

  3. Good job. Just wondering, have you looked into A.I being used for Fusion Research? That is really exciting stuff, and hopefully it’ll lead to us getting our breakthrough in energy.

  4. Great job, Cassie. It is pretty incredible how far we have gotten in technological development. Looks like we have a while to wait before we get personal helper droids.

  5. Elizabeth Webel

    As a close friend of mine says, “AI will be the cause of the pancake apocalypse.”

    But amazing article, good job!!